Improving Efficiency of Data Intensive Applications on GPU Using Lightweight Compression

Advances in Intelligent Systems and Computing

Rzążewski

2015

Self Cite

Abstract. Traversing huge graphs is a crucial part of many real-world problems, including graph databases. We show how to apply Fixed Length lightweight compression method for traversing graphs stored in the GPU global memory. This approach allows for a significant saving of memory space, improves data alignment, cache utilization and, in many cases, also processing speed. We tested our solution against the state-of-the-art implementation of BFS for GPU and obtained very promising results.

Section: Lightweight Compression Methodsmentioning

confidence: 99%

Improving High-Performance GPU Graph Traversal with Compression

Advances in Intelligent Systems and Computing

Rzążewski

2015

Self Cite

“…Copy Serialization Challenge is partially addressed in the set of our publications on lightweight compression methods for GPU environment [12,13,14]. CUDA programming framework allows, in certain circumstances, to perform the computation in parallel to the data transfer.…”

Section: Evaluation Of Challenges Of Heterogeneous (Cpu/gpu) Query Prmentioning

confidence: 99%

“…It may diminish the gain of the acceleration credited to GPU. In our previous works we showed that this data transfer cost can be reduced using lightweight compression methods [12,13,14]. However, this does not solve all the problems since not all tasks may be effectively faster on GPU.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Bi-objective Optimization Framework for Heterogeneous CPU/GPU Query Plans

Fundamenta Informaticae

Stencel

2014

Self Cite

Graphics Processing Units (GPU) have significantly more applications than just rendering images. They are also used in general-purpose computing to solve problems that can benefit from massive parallel processing. However, there are tasks that either hardly suit GPU or fit GPU only partially. The latter class is the focus of this paper. We elaborate on hybrid CPU/GPU computation and build optimization methods that seek the equilibrium between these two computation platforms. The method is based on heuristic search for bi-objective Pareto optimal execution plans in presence of multiple concurrent queries. The underlying model mimics the commodity market where devices are producers and queries are consumers. The value of resources of computing devices is controlled by supply-and-demand laws. Our model of the optimization criteria allows finding solutions of problems not yet addressed in heterogeneous query processing. Furthermore, it also offers lower time complexity and higher accuracy than other methods.

“…This paper describes a dynamic compression strategy planner for time series databases using GPU processors with reasonable processing time and compression ratios. What is even more important, the resulting compressed data block can be decompressed very quickly directly into the GPU memory additionally allowing for ultra fast query processing, what we discussed in our previous publication [12].…”

Section: Introductionmentioning

confidence: 99%

Dynamic Compression Strategy for Time Series Database Using GPU

Advances in Intelligent Systems and Computing

2014

Self Cite

Nowadays, we can observe increasing interest in processing and exploration of time series. Growing volumes of data and needs of efficient processing pushed research in new directions. GPU devices combined with fast compression and decompression algorithms open new horizons for data intensive systems. In this paper we present improved cascaded compression mechanism for time series databases build on Big Table-like solution. We achieved extremely fast compression methods with good compression ratio.